snowflake_tpcds_sales_spoke

A dbt project built on TPC-DS data in Snowflake. This is a downstream "spoke" project that depends on snowflake_tpcds_core via dbt Cloud Mesh.

Local Development

This project uses dbt Cloud Mesh cross-project refs, which don't resolve in dbt-core. A wrapper script handles this so you can run dbt commands locally.

Prerequisites

Install dbt-snowflake in your Python environment:

pip install dbt-snowflake

Running dbt locally

Use the wrapper script instead of calling dbt directly:

scripts/with-local-dbt.sh parse
scripts/with-local-dbt.sh compile --select fct_daily_sales
scripts/with-local-dbt.sh ls --resource-type model
scripts/with-local-dbt.sh test --select stg_tpcds_core__date_dim

The wrapper temporarily injects stub models for the upstream snowflake_tpcds_core project and uses a dummy Snowflake profile. It restores all files when done.

Commands that need a live warehouse connection (dbt run, dbt test on the full project) require real Snowflake credentials in ~/.dbt/profiles.yml and won't work with the dummy profile.

See stubs/README.md for how the stubs work and how to extend them.

Validating the manifest

To validate project fixture completeness (materialization types, resource types, etc.):

# Generate a manifest if you don't have one
scripts/with-local-dbt.sh parse

# Validate
python scripts/verify_manifest.py
python scripts/verify_manifest.py --manifest path/to/other/manifest.json

Resources

• dbt documentation
• dbt Community
• dbt Events
• dbt Blog